ORIGINAL RESEARCH ARTICLEpublished: 14 October 2013

doi: 10.3389/fneur.2013.00154

Primary blast traumatic brain injury in the rat: relatingdiffusion tensor imaging and behaviorMatthew D. Budde1*, Alok Shah1, Michael McCrea1,2,William E. Cullinan3, Frank A. Pintar 1,2 andBrian D. Stemper 1,2

1 Department of Neurosurgery, Medical College of Wisconsin, Milwaukee, WI, USA2 Clement J. Zablocki Veterans Affairs Medical Center, Milwaukee, WI, USA3 Department of Biomedical Sciences, College of Health and Sciences, Marquette University, Milwaukee, WI, USA

Edited by:Firas H. Kobeissy, University ofFlorida, USA

Reviewed by:Joseph Long, Walter Reed ArmyInstitute of Research, USAZhihui Yang, University of Florida, USAKaren M. Von Deneen, XidianUniversity, China

*Correspondence:Matthew D. Budde, Department ofNeurosurgery, VA MedicalCenter-Research 151, Medical Collegeof Wisconsin, 5000 West NationalAve, Milwaukee, WI 53295, USAe-mail: mdbudde@mcw.edu

The incidence of traumatic brain injury (TBI) among military personnel is at its highest pointin U.S. history. Experimental animal models of blast have provided a wealth of insight intoblast injury. The mechanisms of neurotrauma caused by blast, however, are still underdebate. Specifically, it is unclear whether the blast shockwave in the absence of headmotion is sufficient to induce brain trauma. In this study, the consequences of blast injurywere investigated in a rat model of primary blastTBI. Animals were exposed to blast shock-waves with peak reflected overpressures of either 100 or 450 kPa (39 and 110 kPa incidentpressure, respectively) and subsequently underwent a battery of behavioral tests. Diffusiontensor imaging (DTI), a promising method to detect blast injury in humans, was performedon fixed brains to detect and visualize the spatial dependence of blast injury. Blast TBIcaused significant deficits in memory function as evidenced by the Morris Water Maze,but limited emotional deficits as evidenced by the Open FieldTest and Elevated Plus Maze.Fractional anisotropy, a metric derived from DTI, revealed significant brain abnormalitiesin blast-exposed animals. A significant relationship between memory deficits and brainmicrostructure was evident in the hippocampus, consistent with its role in memory func-tion. The results provide fundamental insight into the neurological consequences of blastTBI, including the evolution of injury during the sub-acute phase and the spatially dependentpattern of injury. The relationship between memory dysfunction and microstructural brainabnormalities may provide insight into the persistent cognitive difficulties experienced bysoldiers exposed to blast neurotrauma and may be important to guide therapeutic andrehabilitative efforts.

Keywords: traumatic brain injury, blast neurotrauma, memory dysfunction, diffusion tensor imaging, magneticresonance imaging

INTRODUCTIONThe incidence of traumatic brain injury (TBI) among militarypersonnel in modern combat is the highest of any conflict in U.S.history (1, 2). Up to 20% of combat veterans met the criteriafor TBI on post-deployment screening (3), and exposure to blastsfrom improvised explosive devices contributed to the unprece-dented rate of mild TBI (mTBI). The overwhelming majority ofevents were categorized as mTBI (4), typically characterized bya consistent clinical scenario of postconcussive symptoms, cog-nitive dysfunction, and other functional impairments that followa gradual course of recovery during the initial days and weeksafter injury (5, 6). However, increasing evidence suggests that evenmild head trauma may result in postconcussive symptoms andcomorbidities many months after mTBI (7–10). Moreover, even asingle blast episode may have long-term pathogenic potential andcause enduring neurodegeneration (11). Understanding the neu-rological consequences of blast neurotrauma is essential to enableprevention and guide therapeutic and rehabilitative efforts.

Despite the considerable number of experimental investiga-tions of blast TBI using rodent models reported in recent years,

the mechanisms underlying blast injury and the factors affectingoutcomes remain unclear. Whole-body or head-only blast par-adigms have highlighted the potential for different mechanismsof brain injury following blast exposure (12, 13). Likewise, theeffects of the blast shockwave alone (primary injury) may be dif-ferent than those involving rapid head acceleration subsequent toblast (tertiary injury) and were recently compared directly (14,15). Understanding the time-course of recovery from blast neu-rotrauma is also vital to relate injury mechanisms with patientoutcomes. Behavioral assessments conducted at acute (12, 16–19),sub-acute (20, 21), and chronic time points (13, 20, 21) indicateprolonged neurological abnormalities following blast. However,given the differences in injury protocols and other experimentalvariations between studies conducted at a single timepoint afterinjury, assessments performed at several time points would enablea better characterization of the time-course of deficits.

Another important factor to understand injury tolerance andclinical outcome is the effect of blast magnitude (i.e., shock-wave overpressure). A number of experimental protocols incorpo-rated differing overpressure severities to examine the physiological

www.frontiersin.org October 2013 | Volume 4 | Article 154 | 1

Report Documentation Page Form ApprovedOMB No. 0704-0188

Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering andmaintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, ArlingtonVA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if itdoes not display a currently valid OMB control number.

1. REPORT DATE 14 OCT 2013 2. REPORT TYPE

3. DATES COVERED 00-00-2013 to 00-00-2013

4. TITLE AND SUBTITLE Primary blast traumatic brain injury in the rat: relating diffusion tensorimaging and behavior

5a. CONTRACT NUMBER

5b. GRANT NUMBER

5c. PROGRAM ELEMENT NUMBER

6. AUTHOR(S) 5d. PROJECT NUMBER

5e. TASK NUMBER

5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) Veterans Affairs Medical Center,Milwaukee,WI,53295

8. PERFORMING ORGANIZATIONREPORT NUMBER

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)

11. SPONSOR/MONITOR’S REPORT NUMBER(S)

12. DISTRIBUTION/AVAILABILITY STATEMENT Approved for public release; distribution unlimited

13. SUPPLEMENTARY NOTES

14. ABSTRACT

15. SUBJECT TERMS

16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT Same as

Report (SAR)

18. NUMBEROF PAGES

13

19a. NAME OFRESPONSIBLE PERSON

a. REPORT unclassified

b. ABSTRACT unclassified

c. THIS PAGE unclassified

Standard Form 298 (Rev. 8-98) Prescribed by ANSI Std Z39-18

Budde et al. Blast TBI in the rat

consequences of higher magnitude exposures (22–27), but fewerstudies have assessed whether higher magnitude overpressures leadto greater behavioral deficits. This is important for the quantifi-cation of injury tolerance and may also relate to the durationof the recovery curve. While some studies demonstrated greatermortality rates at higher overpressures (18), a lack of a consistentrelationship between overpressure and cognitive ability (19, 28,29), or neuromotor performance (18) has frequently been noted.A recent study by Cernak et al. demonstrated dose dependency inRotarod and Open Field Test (OFT) outcomes for mice exposedto whole-body 103 and 190 kPa shockwaves (13). However, a needexists to elucidate the dose-response effects and temporal recov-ery of behavioral outcomes following primary blast shockwaveexposure, given somewhat conflicting outcomes from previousinvestigations.

Expanded efforts to screen military personnel for potentialexposure to mTBI have addressed the growing burden of headtrauma in combat. However, there is still no definitive markerof mTBI, which has hampered clinical decision-making. Non-invasive imaging techniques, particularly magnetic resonanceimaging (MRI), have considerably advanced the detection andunderstanding of subtle brain changes following mTBI. Diffusiontensor imaging (DTI), an advanced MRI technique, is a promis-ing method for non-invasively examining the effects of TBI. DTIprobes the molecular motion of water within living tissues toinfer the presence of microscopic structural abnormalities. DTIhas uncovered changes in veterans exposed to blast TBI (30–32)and mTBI through other physical (non-blast) mechanisms (33,34). However, despite its promise, many questions remain regard-ing the use of DTI as a diagnostic marker. While experimentalmodels of blast TBI were essential in identifying the physiological,metabolic, and pathological consequences of blast TBI, DTI hasbeen infrequently applied to experimental animal models of blastunder well-controlled laboratory conditions that would providefundamental insight into the neurotrauma caused by blast forces.

The objective of this work was to examine the behavioralchanges from acute to sub-acute time points, examine the depen-dence of behavioral outcomes on different levels of shockwaveoverpressures, and quantify structural damage to brain tissuesusing DTI following exposure to blast. Importantly, the studyemployed a head-only primary blast injury paradigm to pro-vide fundamental insight into the mechanisms of blast shockwaveneurotrauma and its associated behavioral consequences.

MATERIALS AND METHODSSHOCK TUBE APPARATUSA custom shock tube (Figure 1A) with a 3.6-cm inner diameter,3.0-m driven section, and 0.3-m driver section was used to cre-ate shockwaves with different overpressure magnitudes. A mylarmembrane separated driver and driven sections. The driver sectionwas pressurized with helium until membrane rupture. The shocktube was previously characterized and was shown to produceaccurate and repeatable shockwaves across a range of shockwaveoverpressure magnitudes (35). Pressure transducers (Piezotronics,Inc., Depew, NY, USA) were oriented to record face-on (reflected)pressures at a sampling rate of 5 MHz at selected locations and wereplaced immediately adjacent to the head for all blast exposures.

Additional characterization of the blast pressures was confirmedby placing sensors at identical locations in the absence of a rat,and sensors were oriented to record either face-on (reflected) orside-on (incident) pressures.

ANIMAL PROCEDURESAll animal procedures were approved by the Institutional AnimalCare and Use Committee (IACUC) at our institution. All animalswere allowed access to food and water ad libitum before and aftershockwave exposure. A 2× 2 factorial design was employed withblast magnitude and post-injury duration as independent vari-ables. Animals were exposed to a single blast shockwave with eithera 100 or 450 kPa reflected peak overpressure (Figure 1B), hereafterreferred to the low-blast and high-blast conditions, respectively.Assessments were conducted at acute (1–4 days post injury) orchronic (28–31 days post injury) time points, hereafter referred toas the 4- and 30-days post-injury (DPI) groups, respectively. Eachexperimental group consisted of 9–14 rats. A separate cohort ofeight animals served as sham controls and underwent identicalprocedures without exposure to shockwave.

On the day of injury, Sprague-Dawley rats were anesthetizedwith 4% isoflurane in oxygen and placed in a nose cone for con-tinuous delivery of 1.5% isoflurane. Rats were placed 39 and17 cm from the shock tube opening for the low- and high-blastshockwave exposure conditions, respectively (Figure 1A). Sincethe exhaust gasses, or blast wind, can lead to considerable headacceleration, animals were placed 40° and 20° lateral to the shocktube axis to limit blast wind exposure for the low and high-blastconditions, respectively. Prior characterization of the shock tubedemonstrated minimal blast wind effects at the locations chosenfor this study (35). Moreover, the head was constrained later-ally and inferiorly to prevent head rotational acceleration-inducedinjury, and all shockwave exposures were conducted with the rathead perpendicular to the radial axis from the shock tube opening(Figure 2A). A metal cylinder was also placed around the bodyto limit shockwave overpressure exposure of the torso. Prior workindicated that the cylinder was effective at reducing peak over-pressures to <20% of overpressure magnitudes recorded at thehead. Following shockwave exposure, animals were removed fromanesthesia and allowed to breathe freely. Animals were continu-ously monitored until return of the righting reflex, returned totheir cages for recovery and observation for at least 15 min, andperiodically monitored for 6 h post injury.

Following completion of the behavioral testing protocol, ani-mals were euthanized with an overdose of pentobarbital and per-fused through the left cardiac ventricle with 250 ml of phosphatebuffered saline (PBS) followed by 250 ml of 4% paraformaldehydein PBS. Brains were excised from the skull and placed in fixativefor 48 h followed by PBS for long-term storage.

HIGH-SPEED VIDEOGRAPHYHigh-speed videography was used to record head motion duringshockwave exposure. The digital video camera (Integrated DesignTools Inc., Tallahassee, FL, USA) was placed directly overhead, per-pendicular to the horizontal plane of the animal, to monitor headmotion at 6000 frames per second. Physical markings were placedat the center of the head (i.e., lambda landmark) and between

Frontiers in Neurology | Neurotrauma October 2013 | Volume 4 | Article 154 | 2

Budde et al. Blast TBI in the rat

FIGURE 1 | Blast tube characterization. A shocktube (A) pressurizedwith helium delivered blast shockwaves. Rats were placed at specifieddistances from the shocktube opening based on initial characterizationand were placed off-axis to avoid blast wind. Pressure sensors depict

the typical reflected pressures (B) experienced during the blastshockwave. Blast peak overpressures realized during actual exposuresto the rat head (C) were highly accurate and reproducible (n= 8 foreach).

the eyes (slightly anterior to the bregma). Video tracking software(Tracker 4.75, Open Source Physics) was used to track displace-ment of markings and derive head linear and rotational motiontime histories. Angular position data were digitally filtered usinga lowpass Butterworth filter with a 500-Hz cutoff frequency fol-lowed by numerical differentiation to derive angular velocity andacceleration.

MORRIS WATER MAZEThe Morris water maze (MWM) Visuo-Spatial Learning Para-digm (36, 37) quantified post-traumatic anterograde amnesia andspatial learning. The MWM consisted of a 10 cm diameter plat-form submerged 1 cm below the surface of a 25 cm deep poolof water. Pool diameter was 183 cm. Three testing sessions wereconducted over three consecutive days. Each session consisted ofeight separate trials wherein rats were initially placed at each of thefour cardinal locations (N, E, S, W). During each trial, rats wereallowed to swim until finding and mounting the hidden platformor until 60 s had passed. The location platform was randomizedbetween sessions but was consistent within the same session, andthe first trial of each session was excluded from analysis. Visualcues, including those internal to the maze and external cues in theroom, were maintained between sessions. A computerized trackingsystem and software (Ethovision V8.0, Noldus Information Tech-nology, Wageningen, The Netherlands) recorded movement of therats within the maze. Latency to find the hidden platform and

number of unsuccessful trials were computed from these record-ings. Spatial learning deficits manifest as greater latencies and ahigher number of unsuccessful trials.

ELEVATED PLUS MAZEThe elevated plus maze (EPM) assessment is an ethologically rel-evant test that quantified anxiety-related behaviors. The mazeconsisted of four perpendicular 10 cm× 90 cm arms connectedby a 10 cm× 10 cm central platform. One pair of opposing armswas enclosed by 40-cm high walls, while the other two arms and thecenter platform were uncovered. The entire apparatus was com-posed of black plexiglass and was raised 50 cm above the floor.Rats were initially placed on the central platform facing one of theopen arms and allowed to explore the maze for 5 min. A cameramounted above the maze and Ethovision software automaticallyquantified the number of arm changes, time spent in the openarms, and the number of head dips. Behaviors associated withincreased anxiety included decreased time in the open arms andfewer head dips.

OPEN FIELD TESTThe OFT was used to examine anxiety and spontaneousexploratory behavior. The apparatus consisted of a flat, 100-cmdiameter circular arena with 20-cm walls. Illumination was pro-vided by overhead lights. Rats were initially placed in the centerand allowed to explore the arena for 5 min. The center of the arena

www.frontiersin.org October 2013 | Volume 4 | Article 154 | 3

Budde et al. Blast TBI in the rat

FIGURE 2 | Head kinematics during blast. Blast was directed to the leftside of all animals (A). Landmarks of ink placed between the eyes (red), andon the head (blue) were tracked with high-speed video to obtain headposition (drawn to scale) during exposure to the blast wave, with anexample of the high-blast shown. Angular position (B) was derived byintersection of a line passing through the two points compared to the initialposition of the head, and demonstrated only limited head motion.

consisted of a 66-cm diameter circle, with the remainder consid-ered the border region. The distinction was indicated within thesoftware and did not include any physical markings. An overheadmounted camera along with Ethovision software automaticallyquantified measures including total distance traveled and totaltime spent in the center of the maze. Behaviors associated withincreased anxiety include a decreased time spent in the center.

STATISTICAL ANALYSISA two-factor Analysis of Variance (ANOVA) was used to identifystatistically significant interactions for behavioral metrics betweenblast severity and post-injury time for groups exposed to blast,followed by post hoc tests when warranted. In the absence of asignificant interaction, a one-way ANOVA was conducted withthe sham group included to test for a significant main effect ofeither blast magnitude or time after injury using separate mod-els. Post hoc comparisons were performed if the overall test wassignificant. For all tests, significance was set at a p-value of <0.05.

MAGNETIC RESONANCE IMAGINGA 9.4 T horizontal bore Bruker BioSpec was employed forMRI procedures. Fixed brains were immersed in susceptibility-matching fluid (Fomblin; Solvay Solexis, NJ, USA) and placedin a custom 20-mm diameter inductively coupled loop-gap coil.

Room temperature was maintained throughout the experiment.A series of gradient echo scout images were acquired to ensurereproducible slice orientations and placement within the magnet.A spin-echo sequence (TR= 2000 ms, TE= 21 ms) with Stejskal–Tanner pulsed gradients was used to perform diffusion-weightingalong 12 non-orthogonal directions at a b-value of 1200 s/mm2

and two non-diffusion weighted images (b= 0 s/mm2). A double-echo spin-echo readout (echo spacing= 3.3 ms) was used toimprove signal to noise ratio (SNR) by magnitude averaging ofthe individual echo images. Twenty axial slices at a thickness of0.5 mm and an in-plane resolution of 0.2 mm2 covered the entirebrain and brainstem. The entire MRI experiment was completedin approximately 1 h.

MRI DATA ANALYSISDiffusion weighted images were corrected for eddy current distor-tions, and the diffusion tensor was calculated on voxel-by-voxelbasis using fMRI Software Library (FSL). For spatial registration, astudy-specific template was created using tensor-based registrationimplemented in the DTI-TK software package (38). An iterativeprocedure using rigid body, affine, and diffeomorphic registra-tion was used to create a template at a final resolution of 200 µm3.Tensor volumes were spatially smoothed with an anisotropic three-dimensional filter implemented in Matlab (39). Maps of fractionalanisotropy (FA) were derived and used for subsequent statisticalanalysis.

Differences between blast-exposed groups and the sham groupwere assessed using a two-sample t -test. Permutation-based infer-ence testing implemented in the FSL was used for statisticalhypothesis testing, and voxels with p < 0.05 (uncorrected formultiple comparisons) and clusters >100 voxels were consideredstatistically significant. Relationships between FA and behavioraltests were examined using linear regression analysis of the blast-exposed animals. For each of the behavioral tests (MWM, EPM,and OFT), behavioral outcomes were regressed against FA val-ues on a voxel-by-voxel basis to identify voxels with a significantF-statistic (p < 0.05).

A complementary region of interest (ROI) analysis was per-formed to assess distributed FA abnormalities. The mean FA mapfrom all animals (n= 60) was segmented into gray and white mat-ter using a threshold of 0.25. The ROIs were manually edited toinclude only the cerebral white matter tracts and cerebral cortexby excluding the cerebellum, brainstem, and subcortical structures,and each hemisphere was assessed separately. The FA map fromeach animal was converted to z-scores (z = (x−µ)/σ) on a voxel-by-voxel basis where x is the individual FA value, µ and σ are themean and standard deviation FA value of the sham group, respec-tively. To ensure robustness, an additional six blast sham animalswere included (n= 14). Moreover, a leave-one-out approach wasused for the sham group by calculating the z-scores for each of theindividual brains compared to the mean and standard deviation ofthe remaining 13. The percentage of voxels exceeding a z-score of±2 was calculated separately for the WM and GM ROIs (31, 34).The resulting quantitative metric, percentage of abnormal FA vox-els, was compared to the sham group for each of the blast-exposedgroups using a Student’s t -test and a Bonferroni correction formultiple comparisons.

Frontiers in Neurology | Neurotrauma October 2013 | Volume 4 | Article 154 | 4

Budde et al. Blast TBI in the rat

HISTOLOGYFixed brains were processed for routine paraffin embedding andcut at a thickness of 4 µm. Selected sections were stained in 0.01%toluidine blue O for 5 min and mounted. Alternatively, sectionsfor immunofluorescent staining were placed in 10 mM sodiumcitrate buffer in PBS, microwaved for 5 min, and allowed to coolto room temperature. Sections were incubated at 4°C overnightwith primary antibodies for glial fibrillary acidic protein (GFAP;1:500; EMD Millipore, Billerica, MA, USA) or cleaved caspase-3 (1:500, EMD Millipore, Billerica, MA, USA). Incubation withthe secondary antibody (Alexa Fluor; 1:500, Molecular Probes)was performed for 30 min at room temperature. Sections weremounted and imaged with a Nikon microscope.

RESULTSBLAST SHOCKWAVE AND TBI MODEL CHARACTERIZATIONPressure time histories depict the standard experimental approx-imation of the Friedlander waveform of the primary shockwaveoverpressure (Figure 1B). Peak reflected overpressures measuredat the rat head had a high accuracy and reproducibility, with mea-sured values of 451± 25 (n= 21) and 112± 14 (n= 22) kPa forthe 450 and 100 kPa conditions, respectively (Figure 1C). Posi-tive durations were 0.34± 0.15 ms for the high-blast exposuresand 0.46± 0.21 ms for the low-blast exposures. Pressures mea-sured near the rat thorax were considerably lower at 63± 26 and19± 5 kPa, respectively. To ensure the presence of the rat didnot alter blast pressure measurements, separate blasts were per-formed with sensors placed in the location of the rat head butwithout a rat in position. Measurements demonstrated similarreflected pressures of 450± 23 kPa (n= 3) and incident pressuresof 110± 5.6 kPa (n= 6). The experimental apparatus also limitedhead motion. Using high-speed video analysis, a maximum lateraldisplacement of 6.98 mm was measured in the high-blast condi-tion, with an angular acceleration of <150 krad/s2 (Figure 2B).Minimal head motion was noted for the low-blast condition withan acceleration of <10 krad/s2. Therefore, any behavioral deficitsor structural brain damage resulting from these exposures wasmost likely the result of shockwave passage through the braintissues and its subsequent effects rather than head rotationalacceleration.

MEMORY DEFICITS IN SHOCKWAVE EXPOSURE GROUPSOn the MWM Visuo-Spatial Learning Paradigm, all experi-mental groups demonstrated the ability to learn (Figure 3A),since latency to find a hidden platform decreased over time(p < 0.001). On the first day of testing, there were no sig-nificant differences in latency between the groups (p= 0.18).On the second day, a main effect of blast severity approached,but did not meet the threshold for significance (p= 0.06). Onthe third day, the main effect of blast severity was signifi-cant (p= 0.034). Both the low-blast (p= 0.015) and high-blast(p= 0.016) groups had significantly greater latencies to find theplatform than the sham group (Figure 3B), but the two groupswere not significantly different than one another (p= 0.97). Timeafter blast had no effect on latency between the 4 and 30 DPIgroups (p= 0.86), indicating persistent cognitive deficits followingshockwave exposure.

FIGURE 3 | Morris water maze. Memory and learning deficits wereevident in the blast-exposed animals compared to sham animals (A). Swimpath traces from the final trial (B), with each animal shown in a differentcolor, reveal the difficulties in finding the hidden platform (gray) in theblast-exposed animals compared to the sham animals. The entry point isindicated with an arrow. Error bars indicate SEM. *p < 0.05.

The number of unsuccessful trials across all 18 sessions of theMWM also indicated impaired memory and cognition in animalsexposed to blast (Figure 3A). There was no significant interac-tion between blast severity and time (p= 0.73), but a significantmain effect of blast severity was evident (p < 0.019). The high-blast group had significantly more unsuccessful trials comparedto the sham blast group (p= 0.006), whereas the low-blast groupwas not significantly different from animals exposed to either high-blast (p= 0.075) or sham (p= 0.15) blast. The main effect of timeafter injury for the number of unsuccessful trials was not signifi-cant (p= 0.18). Time spent in the target quadrant had a significantinteraction between blast severity and time after injury (p= 0.04).In the post hoc analysis, the low-blast group at 30 DPI spent sig-nificantly more time in the target quadrant than the high-blastgroups at either 4 (p= 0.029) or 30 (p= 0.01) DPI. No otherindividual groups were significantly different from one another. Amain effect of blast severity was also significant (p= 0.01), with thelow-blast group spending more time in the target quadrant thanthe high-blast (p= 0.008) or sham (p= 0.031) groups. All othercomparisons for time in the target quadrant were not significant.

OPEN FIELD ASSESSMENTThe OFT was used to assess locomotor activity and anxiety behav-ior (Figure 4A). Activity, as measured by the total distance traveledduring the test, did not have a significant interaction betweenblast severity and time after injury (p= 0.69). A significant main

www.frontiersin.org October 2013 | Volume 4 | Article 154 | 5

Budde et al. Blast TBI in the rat

FIGURE 4 | Open field test and elevated plus maze. Emotionality andanxiety was examined with the Open Field Test (A) and the Elevated PlusMaze (B) for each of the four blast conditions and sham group. Error barsindicate SEM. *p < 0.05.

effect of blast severity was observed (p= 0.029). Compared to thesham blast group, animals exposed to low (p= 0.018) or high(p= 0.011) blast conditions had significantly less activity, but theywere not significantly different from one another (p= 0.76). Theeffect of time after injury was not significant (p= 0.84). Likewise,the amount of time spent in the center of the OFT, a measure ofanxiety, did not have a significant interaction between blast sever-ity and time after injury (p= 0.20). Time in the center was also notsignificant for either the main effect of blast severity (p= 0.51) ortime after injury (p= 0.63).

EMOTIONALITY CHANGES IN SHOCKWAVE EXPOSURE GROUPSExploratory behavior on the EPM served as an indicator of emo-tionality. The amount of time spent in the open arms of the EPM(Figure 4B), a metric associated with anxiety, did not have a sig-nificant interaction with blast severity and timing after injury(p= 0.39). However, the main effect of blast severity was signifi-cant (p= 0.037). The high-blast group spent significantly less timein the open arms compared to the low-blast (p= 0.013) group. Thehigh-blast (p= 0.11) and low-blast (p= 0.71) groups were not sig-nificantly different from the sham group. The main effect of timeafter injury was not significant (p= 0.19).

The number of arm changes in the EPM did not have a sig-nificant interaction between blast severity and time after injury

(p= 0.63). The main effect of blast severity was not significant(p= 0.35), although the number of arm changes was decreasedby 26 and 29% in the low and high-blast groups, respectively(Figure 4B). The main effect of timing after injury was notsignificant (p= 0.98).

SHOCKWAVE EXPOSURE CAUSED SPATIALLY DEPENDENT ANDPROGRESSIVE DTI ABNORMALITIESShockwave exposure caused significant changes in FA. A distinctpattern of abnormalities emerged related to blast severity and timeafter injury, with significant differences between each of the blast-exposed groups compared to the sham group (Figure 5A). At 4DPI, significant decreases in FA were evident in the ipsilateral cor-tex, medial prefrontal cortex (mPFC), hippocampus, and thalamusand were more pronounced in the animals exposed to high-blastthan those exposed to low-blast shockwaves. At 30 DPI, significantdecreases in FA were evident in both blast exposures. Brain regionsaffected at 30 DPI included those affected at 4 DPI along withadditional injury evident in the contralateral cortex and brain-stem. The voxelwise statistical tests were uncorrected for multiplecomparisons to aid in visualizing the spatial pattern of FA changesacross all of the blast-exposed groups with the same sham groupused for all comparisons. The decreases in FA were greater andmore widespread in the group exposed to high-blast comparedto those exposed to low-blast. The three-dimensional visualiza-tion of significant differences are shown in Figure 5B. The effectsizes (Cohen’s D) for the significant clusters in each of the groupscompared to the sham group were 0.417 and 0.407 for the low-blast and high-blast at 4 DPI, respectively, and 0.493 and 0.495for the low-blast and high-blast groups, respectively, at 30 DPI.When corrected for multiple comparisons at p < 0.05 by control-ling the familywise error, the voxelwise differences did not surviveand were therefore followed up by an unbiased ROI analysis.

A ROI analysis was performed to complement the voxelwisefindings and quantitatively examine the laterality of the DTIchanges (Figure 6). ROIs encompassing the left (ipsilateral) orright (contralateral) cerebral white matter (Figure 6A) or corticalgray matter (Figure 6B) revealed that mean FA was not signifi-cantly different between any of the groups. However, an alternativemethod quantifying the number of abnormal FA voxels for eachanimal relative to a spatially aligned sham group (z-score) revealedsignificant differences between groups. In the white matter, thepercentage of abnormal (|z | > 2) voxels was significantly greater inall of the blast-exposed groups compared to shams in the ipsilateralwhite matter, with greater changes observed at 30 DPI (Figure 6C).In the contralateral white matter, no significant differences wereobserved. However, in the cerebral cortex (Figure 6D), the numberof abnormal FA voxels was significantly greater in the ipsilateralcortex for all blast-exposed groups compared to the sham group(all p < 0.05, corrected for multiple comparisons). In the con-tralateral hemisphere, the low-blast (p= 0.0009) and high-blast(p= 0.0005) exposure groups at the chronic time points were sig-nificantly different than the sham group, whereas no significantdifferent differences were noted at the acute time point. In general,the ROI results confirmed the previous voxelwise observations. Toidentify whether the FA changes were predominantly associated

Frontiers in Neurology | Neurotrauma October 2013 | Volume 4 | Article 154 | 6

Budde et al. Blast TBI in the rat

FIGURE 5 | Microstructural abnormalities following primary blast. Thebrains of animals exposed to blast shockwave underwent DTI and comparedto animals exposed to sham blast (A). Significant FA decreases were evidentat 4 (left) and 30 days post-blast (right). At both timepoints, greater numbersof significant voxels were evident in the animals exposed to high-blastcompared to those at exposed to low-blast. Compared to 4 days post-blast,

there were more significant voxels at 30 days post-blast in both blastexposures. Decreased FA was most prominent in the cortex, but alsoincluded the thalamus, and ipsilateral ventral hippocampus. Thethree-dimensional distribution of significant changes compared to sham areshown (B) as viewed from the top of the brain. The direction of blast isindicated by the arrow and is identical for both panels.

with changes in axial diffusivity or radial diffusivity, these valueswere obtained from the abnormal FA voxels from each animalsand averaged across groups. There were no significant differencesin AD or RD between any of the groups compared to the shamgroup.

RELATIONSHIP BETWEEN FRACTIONAL ANISOTROPY AND BEHAVIORLinear regression analysis was performed to identify relationshipsbetween behavioral outcomes and microstructure assessed withDTI (Figure 7). In the MWM, clusters of voxels with significantlinear relationships between FA and behavioral outcomes wereevident bilaterally in the hippocampus, predominantly in the dor-sal region. A significant region was also evident in the motorcortex, predominantly in the contralateral cortex. Across all sig-nificant voxels, target quadrant percentage was a better predictorof FA (p < 0.001) than number of unsuccessful trials (p= 0.045)or latency (p= 0.088). In the OFT, outcomes were significantlyrelated to FA of the contralateral, ventral hippocampus, and a smallportion of the ipsilateral external capsule. FA in these regions wassignificantly associated with both outcomes of the OFT, time in thecenter (p < 0.001), and distance traveled (p < 0.001). In the EPM,only a small region of significance was located in the contralateralcortex. Neither the number of arm changes (p= 0.082) or timespent in the open arms (p= 0.30) were independent predictors ofFA in this region.

HISTOLOGICAL EVIDENCE OF INJURYNo hemorrhage was evident in any of the animals exposed toshockwaves based on visual inspection of the brain surface. A

qualitative analysis revealed greater astrocyte hypertrophy in themPFC (Figure 8), a region exhibiting significant DTI abnor-malities (Figure 4A), following high-blast shockwave exposurecompared to low-blast shockwave exposure. A similarly greaterhypertrophy was evident at chronic compared to acute timepoints. Correspondingly, sections stained with Toluidine bluerevealed chromalytic neurons in the cortex of brains exposed tohigh-blast shockwaves at both acute and chronic time points,whereas the brain exposed to low-blast shockwaves containedonly sparse darkly stained neurons. Staining for cleaved caspase-3, a nuclear marker of apoptosis, revealed numerous apop-totic cells within the ipsilateral cortex, whereas no apoptoticnuclei were evident in the contralateral cortex or at 30 dayspost-blast.

DISCUSSIONAnimal brain injury models have proven essential to understandthe consequences of blast injury. Although the precise mechanismsof blast neurotrauma are the subject of considerable investigationand debate, several important experimental factors may have aconsiderable influence on the findings. The current investigationemployed an experimental rodent model designed to be a trans-lational paradigm representative of human blast injury dynamics.In our animal model, the body of each rat was protected, so onlythe head was exposed to shockwaves. This protective approach isconsistent with the setting of ballistic personal protective bodyarmor worn by military personnel being effective at attenuat-ing blast overpressure exposure to the thorax (40). Field injuriesare more likely attributable to shockwave exposure of the head,

www.frontiersin.org October 2013 | Volume 4 | Article 154 | 7

Budde et al. Blast TBI in the rat

FIGURE 6 | Region of interest DTI analysis. The mean FA of the whitematter tracts (A) or cerebral cortex (B), assessed separately for theipsilateral and contralateral hemispheres, were not significantlydifferent between any of the groups. In contrast, the percentage ofvoxels with FA values >2 standard deviations from the mean (z -scores)compared to the sham group revealed significant group differences. Inthe white matter (C) no significant differences were evident in either

the ipsilateral or contralateral hemispheres. In the cerebral cortex(D), the percentage of abnormal voxels ipsilateral to the blast exposurewas significantly greater in all blast-exposed groups compared to thesham group. In the contralateral hemisphere, only groups 30 dayspost-blast had a significantly greater percentage of abnormal FA voxelcompared to the sham group. Error bars indicate standarddeviation.*p < 0.05 compared to sham.

and injury tolerances are different between shockwave exposuresto the head and brain tissue compared to those involving pul-monary injury. The scaling of shockwave characteristics from therat to the human is another important experimental concern.Construction of explosive devices using artillery shells (41) hasdemonstrated overpressure durations typically <10 ms. Scalingratios presented by Bass et al. (42), albeit for pulmonary expo-sures, can be used to relate overpressure durations from rodentmodels to human-equivalent values. In the current study, scaledoverpressure durations were consistent with human exposures of2–3 ms. Isolating the effects of blast shockwave from head accel-eration is an important concept to understand the mechanismsof blast injury. In the current study, the head was laterally sup-ported during shockwave exposure, which prevented appreciablehead excursion during the low-blast exposures and limited excur-sions during high-blast exposures. Angular accelerations were wellbelow the threshold for mild concussion in the rat (43). Similarly,positioning rats outside and off-axis from the shock tube limitedexposure to venting gas (i.e., blast wind). While exposure outsideof the tube introduces shockwave dynamics such as diffractionand weakening, all rats were exposed to a repeatable shockwaveprofile (Figure 1) without blast wind confirmed by measured pres-sure profiles. Collectively, the resulting behavioral, imaging, andhistological consequences of blast injury were attributable to theprimary shockwave exposure and not pulmonary or rotationalacceleration mechanisms.

Primary blast injury caused significant memory deficits inshockwave-exposed rats compared to sham-exposed animals, withmore pronounced deficits in the higher shockwave exposuregroup. These data provide preliminary support for the theoret-ical assumption of a “dose effect” associated with more severeor repetitive blast TBI exposure. The observed memory andcognitive deficits were generally consistent with findings fromother experimental investigations of shockwave-induced TBI inrodents (21, 44), although the magnitude of blast exposure dif-ferences between the studies may have implications regardinginjury tolerance to blast. Goldstein and colleagues (14) demon-strated significant cognitive deficits after 77 kPa exposures, but thehigh magnitude head rotational accelerations potentially causedinjury through various mechanisms. The cognitive deficits inthe present protocol, which incorporated head restraint, weredirectly attributable to the interaction of the shockwave andthe head. A notable difference between the present study andprevious investigations is the presence of significant cognitivedeficits acutely after injury. Increased latency to find the plat-form and greater numbers of unsuccessful trials were evident forshockwave-exposed groups at <4 days post injury. Other studieshave reported no differences between controls and shockwave-exposed rats at 3–8 days post injury (21, 44, 45). The disparitycould be related to the use of the Barnes Maze assessment inother studies, compared to the use of the Morris Water Maze inthe current study, since the two assessments may have different

Frontiers in Neurology | Neurotrauma October 2013 | Volume 4 | Article 154 | 8

Budde et al. Blast TBI in the rat

FIGURE 7 | Diffusion tensor imaging correlations with behavior. Acrossall animals exposed to blast, multiple linear regression was used todetermine whether behavioral outcomes were associated with brainmicrostructure. For each behavioral test, the outcome measures obtainedfrom it were regressed against the FA value from all animals using a F -test,controlling for blast severity and DPI. In the MWM, FA was significantlyrelated to behavioral performance in the hippocampus and portions of themotor cortices. In the OFT, a region of the ventral hippocampus and regionof the external capsule were significantly related to the behavioralperformance. The EPM was related to FA in only a small region of thecontralateral cortex.

sensitivities. These differences may also be attributable to differ-ences in the injury protocol such as the used of head restraints andtorso protection.

Neurobehavioral and emotional changes assessed using theEPM and OFT were also generally consistent with previous reports(44). Elevated levels of anxiety were evident for shockwave-exposed rats, with some indication of a dose response and transientsymptoms. Sham rats demonstrated the greatest magnitude ofopen arm time, with the groups exposed to blast spending progres-sively less time in the open arms at the acute time point. Likewise,open field center time decreased in the sham group compared tothe groups exposed to blast. These findings indicate higher levelsof anxiety in more severely injured rats, again supporting a doseeffect of blast TBI. This type of dose dependency for emotionalchanges following shockwave injury may have clinical implica-tions for the understanding of patient condition and outcomes.Whether emotional changes are transient or persistent follow-ing shockwave exposure has been mixed. In the current study,the time in the center of the open field was decreased in thehigh-blast group acutely, but recovered by 1 month post injury.However, open arm time in EPM decreased acutely in the high-blast group and did not resolve by 1 month post injury. Similarbehavioral paradigms at other laboratories have observed pro-gressive emotional changes at 44 DPI (20), but no deficits at66 DPI. Conversely, persistent decreases in open arm time from

48 h to 1 month post injury have also been observed (44). Emo-tional deficits have also been shown to persist up to 6 monthspost-injury (46). The disparities may be related to the interactionbetween stress and emotional dysfunction (44, 47) and requirefurther investigation, particularly in relation to the developmentof PTSD in military personnel and returning veterans exposed toblast during combat (2, 4, 7).

The finding of DTI abnormalities and their spatial distribu-tion provides some insight into the mechanisms of injury fol-lowing shockwave exposure. At acute time points, decreased FAin the superficial cortex ipsilateral to the shockwave suggests alocal effect of the shockwave on the brain tissue, and is consis-tent with previous histological studies demonstrating gliosis inthe lateral (14) and medial prefrontal cortices (44). The effectsmay be related to high shear strain or microcavitation at theinterface between the brain and cerebrospinal fluid shown insimulations of blasts in the human (48) and rat brain (27). Inthe current study, injuries progressed from acute to chronic timepoints and involved greater regions of the cortex as well as por-tions of the brainstem, suggesting an ongoing response similarto that observed elsewhere (49). The injury in the superficialcortex could potentially be related to the acute disruption ofthe blood-brain barrier (17) and it sequelae. Indeed, gliosis is aprominent pathology after blast (15) and hypertrophic astrocyteswere evident in the mPFC in the current study (Figure 7). Whitematter injury was also evident in the current study (Figure 4),consistent with previous histological reports of axonal injury inthe brainstem and cerebellum (50). Furthermore, the regions ofinjury could be associated with auditory regions in the brain(51). Further studies will be necessary to understand the com-plex relationship between biomechanical and other consequencesof blast injury.

The current study employed voxelwise analysis of FA imagessince the orientation of the blast was consistent across all ani-mals. However, since blast neurotrauma in humans is highlyvariable with respect to intensity, direction, and other factors,alternative approaches to capture the heterogeneity have beenproposed. Comparing an individual to a control group to derivegroup-averaged measures (31, 34) or single subject maps (52) ofabnormal FA was shown to have greater sensitivity to TBI thanmeasures of mean FA. In the current study, this approach was usedto quantify the voxelwise changes. As expected, there were no dif-ferences in whole brain mean FA values between any of the groups.However, quantifying the number of abnormal FA voxels in eachsubject relative to the control group and group-averaging this met-ric was sensitive to group differences (Figure 6). The results werelargely consistent with the voxelwise findings, but also providegreater insight into the directionality of the FA changes, since thisis conflicting in human DTI studies following TBI. In the cortex,the FA decrease was greater with higher blast intensity and withlonger durations from the injury. In contrast, FA increased in thewhite matter acutely and appeared to normalize at the 30 day time-point. This is the opposite that of the white matter FA decreasesobserved in severe TBI (53). In human mTBI, however, increasedwhite matter FA was often observed in the acute (54) and semi-acute (34) periods whereas decreased FA was consistently observed

www.frontiersin.org October 2013 | Volume 4 | Article 154 | 9

Budde et al. Blast TBI in the rat

FIGURE 8 | Histological evidence of blast effects. Brain sections from theregions exhibiting DTI changes, approximately +2.0 mm Bregma, werestained with toluidine blue (A). Sections revealed numerous “dark” neuronsin the brains exposed to high-blast (C) at both 4 and 30 days post-blast,particularly in the ipsilateral cortex, but fewer in the contralateral cortex. Onlysporadic abnormal neurons were observed in the low-blast and sham brains.In a brain exposed to high-blast, staining for cleaved caspase-3 (B) was

observed in the ipsilateral cortex at 4 days post-blast, but not in thecontralateral cortex, and the stain was no longer present at 30 days post-blast.No staining was present in the sham brain. In the medial prefrontal cortex, aregion with significant FA changes, astrocytes stained with GFAP exhibitedextensive hypertrophy in the high-blast brain at 30 days compared to thesham brain, but astrogliosis was noticeably less at 4 days post-blast or in thelow-blast condition.

in the more chronic phases (30–32). Thus, although FA is a sen-sitive biomarker of injury following TBI, interpreting the changesin FA are complex with respect to the underlying pathology (55).

The relationship between DTI, cognitive, and neurobehavioralabnormalities identifies a putative link between regional brainabnormalities and the resulting clinical manifestations of blastinjury. While the current findings were observed in an experi-mental animal model, they likely have significant translationalrelevance to our understanding of the pathophysiology and clinicaleffects of blast injury in human. Impairment in memory functionwas related to abnormalities in the hippocampus (Figure 7), asexpected although the precise mechanism for the changes is notclear. A quantitative assessment of the biological basis of the DTIfindings would provide greater insight into the injuries evident inhumans after blast. One limitation of the current study is the useof fixed tissues. Fixation has been shown to maintain anisotropy(56), but in vivo measurements would enable a better appreci-ation of longitudinal changes following TBI. On the other hand,imaging of fixed tissues affords substantial improvements in imagequality and resolution compared to in vivo and thereby better sen-sitivity to subtle changes. Another limitation is that blast injurieswere performed at a single orientation relative to the shock tube(side-on). The importance of orientation has been demonstratedusing simulations (27, 48) and experimental studies (19, 57), but awhole brain assessment such as that provided by DTI has yet to be

performed. Utilization of serial, in vivo MRI will be important tofully characterize the effects of blast TBI and the evolution fromacute to chronic injury.

CONCLUSIONIn a rat model of blast TBI, primary blast exposure was sufficient toinduce cognitive and neurobehavioral deficits and microstructuralinjury detected with DTI. Importantly, both outcomes scaled withthe magnitude of the blast shockwave, reflecting a dose effect thathas implications to the fundamental pathophysiology and neuro-logic sequelae of blast TBI in humans. Moreover, the correlationbetween neurological deficits and DTI findings underscores theconnection between brain injury and behavior that has particularrelevance to advancing the science of therapeutic interventions andprevention of long-term effects of blast injury in military servicemembers and veterans.

ACKNOWLEDGMENTSThis project was partially funded through a grant from the Depart-ment of Veterans Affairs, Veterans Health Administration, Reha-bilitation Research and Development Service, the Department ofNeurosurgery, and the Research and Education Initiative Fund,a component of the Advancing a Healthier Wisconsin endow-ment at the Medical College of Wisconsin (5520207 to MatthewD. Budde).

Frontiers in Neurology | Neurotrauma October 2013 | Volume 4 | Article 154 | 10

Budde et al. Blast TBI in the rat

REFERENCES1. Okie S. Traumatic brain injury

in the war zone. N Engl J Med(2005) 352:2043–7. doi:10.1056/NEJMp058102

2. Warden D. Military TBI dur-ing the Iraq and Afghanistanwars. J Head Trauma Rehabil(2006) 21:398–402. doi:10.1097/00001199-200609000-00004

3. Holdeman TC. Invisible woundsof war: psychological and cognitiveinjuries, their consequences, andservices to assist recovery. Psychi-atr Serv (2009) 60:273–273. doi:10.1176/appi.ps.60.2.273

4. Hoge CW, McGurk D, Thomas JL,Cox AL, Engel CC, Castro CA. Mildtraumatic brain injury in U.S. sol-diers returning from Iraq. N EnglJ Med (2008) 358:453–63. doi:10.1056/NEJMoa072972

5. McCrea M, Guskiewicz KM, Mar-shall SW, Barr W, Randolph C,Cantu RC, et al. Acute effects andrecovery time following concus-sion in collegiate football players:the NCAA concussion study. JAMA(2003) 290:2556–63. doi:10.1001/jama.290.19.2556

6. DePalma RG, Burris DG, ChampionHR, Hodgson MJ. Blast injuries. NEngl J Med (2005) 352:1335–42. doi:10.1056/NEJMra042083

7. Trudeau DL, Anderson J, HansenLM, Shagalov DN, Schmoller J,Nugent S, et al. Findings of mildtraumatic brain injury in combatveterans with PTSD and a history ofblast concussion. J NeuropsychiatryClin Neurosci (1998) 10:308–13.

8. Guskiewicz KM, Marshall SW,Bailes J, McCrea M, Cantu RC,Randolph C, et al. Associationbetween recurrent concussion andlate-life cognitive impairment inretired professional football players.Neurosurgery (2005) 57:719–26.doi:10.1227/01.NEU.0000175725.75780.DD discussion 719–726

9. Kerr ZY, Marshall SW, HardingHP Jr, Guskiewicz KM. Nine-year risk of depression diagno-sis increases with increasing self-reported concussions in retired pro-fessional football players. Am JSports Med (2012) 40:2206–12. doi:10.1177/0363546512456193

10. Verfaellie M, Lafleche G, SpiroA III, Tun C, Bousquet K.Chronic postconcussion symp-toms and functional outcomesin OEF/OIF veterans with self-report of blast exposure. J IntNeuropsychol Soc (2013) 19:1–10.doi:10.1017/S1355617712000902

11. McKee AC, Stein TD, Nowinski CJ,Stern RA, Daneshvar DH, Alvarez

VE, et al. The spectrum of diseasein chronic traumatic encephalopa-thy. Brain (2013) 136:43–64.

12. Long JB, Bentley TL, Wessner KA,Cerone C, Sweeney S, Bauman RA.Blast overpressure in rats: recreatinga battlefield injury in the laboratory.J Neurotrauma (2009) 26:827–40.doi:10.1089/neu.2008.0748

13. Cernak I, Merkle AC, KoliatsosVE, Bilik JM, Luong QT, MahotaTM, et al. The pathobiology ofblast injuries and blast-inducedneurotrauma as identified using anew experimental model of injuryin mice. Neurobiol Dis (2011)41:538–51. doi:10.1016/j.nbd.2010.10.025

14. Goldstein LE, Fisher AM, TaggeCA, Zhang XL, Velisek L, Sulli-van JA, et al. Chronic traumaticencephalopathy in blast-exposedmilitary veterans and a blast neu-rotrauma mouse model. Sci TranslMed (2012) 4:134ra160. doi:10.1126/scitranslmed.3003716

15. Svetlov SI, Prima V, Glushakova O,Svetlov A, Kirk DR, Gutierrez H, etal. Neuro-glial and systemic mecha-nisms of pathological responses inrat models of primary blast over-pressure compared to “composite”blast. Front Neurol (2012) 3:15. doi:10.3389/fneur.2012.00015

16. Cernak I, Wang Z, Jiang J, Bian X,Savic J. Ultrastructural and func-tional characteristics of blast injury-induced neurotrauma. J Trauma(2001) 50:695–706. doi:10.1097/00005373-200104000-00017

17. Readnower RD, Chavko M, AdeebS, Conroy MD, Pauly JR, McCar-ron RM, et al. Increase in blood-brain barrier permeability, oxida-tive stress, and activated microgliain a rat model of blast-induced trau-matic brain injury. J Neurosci Res(2010) 88:3530–9. doi:10.1002/jnr.22510

18. Wang Y, Wei Y, Oguntayo S, WilkinsW, Arun P, Valiyaveettil M, et al.Tightly coupled repetitive blast-induced traumatic brain injury:development and characterizationin mice. J Neurotrauma (2011)28:2171–83. doi:10.1089/neu.2011.1990

19. Ahlers ST, Vasserman-Stokes E,Shaughness MC, Hall AA, ShearDA, Chavko M, et al. Assessmentof the effects of acute and repeatedexposure to blast overpressure inrodents: toward a greater under-standing of blast and the poten-tial ramifications for injury inhumans exposed to blast. Front Neu-rol (2012) 3:32. doi:10.3389/fneur.2012.00032

20. Kovesdi E, Gyorgy AB, Kwon SK,Wingo DL, Kamnaksh A, Long JB,et al. The effect of enriched envi-ronment on the outcome of trau-matic brain injury; a behavioral,proteomics, and histological study.Front Neurosci (2011) 5:42. doi:10.3389/fnins.2011.00042

21. Kovesdi E, Kamnaksh A, WingoD, Ahmed F, Grunberg NE, LongJB, et al. Acute minocycline treat-ment mitigates the symptoms ofmild blast-induced traumatic braininjury. Front Neurol (2012) 3:111.doi:10.3389/fneur.2012.00111

22. Nakagawa A, Fujimura M, KatoK, Okuyama H, Hashimoto T,Takayama K, et al. Shock wave-induced brain injury in rat: noveltraumatic brain injury animalmodel. Acta Neurochir Suppl (2008)102:421–4. doi:10.1007/978-3-211-85578-2_82

23. Bolander R, Mathie B, Bir C, RitzelD, Vandevord P. Skull flexure as acontributing factor in the mech-anism of injury in the rat whenexposed to a shock wave. Ann Bio-med Eng (2011) 39:2550–9. doi:10.1007/s10439-011-0343-0

24. Cullen DK, Browne KD, Xu Y,Adeeb S, Wolf JA, McCarron RM,et al. Blast-induced color changein photonic crystals correspondswith brain pathology. J Neuro-trauma (2011) 28:2307–18. doi:10.1089/neu.2011.1718

25. Bir C, Vandevord P, Shen Y, RazaW, Haacke EM. Effects of variableblast pressures on blood flow andoxygen saturation in rat brain asevidenced using MRI. Magn ResonImaging (2012) 30:527–34. doi:10.1016/j.mri.2011.12.003

26. Risling M, Davidsson J. Experimen-tal animal models for studies on themechanisms of blast-induced neu-rotrauma. Front Neurol (2012) 3:30.doi:10.3389/fneur.2012.00030

27. Sundaramurthy A, Alai A, Gan-pule S, Holmberg A, Plougonven E,Chandra N. Blast-induced biome-chanical loading of the rat: an exper-imental and anatomically accuratecomputational blast injury model.J Neurotrauma (2012) 29:2352–64.doi:10.1089/neu.2012.2413

28. Saljo A, Bolouri H, Mayorga M,Svensson B, Hamberger A. Low-level blast raises intracranial pres-sure and impairs cognitive func-tion in rats: prophylaxis withprocessed cereal feed. J Neuro-trauma (2010) 27:383–9. doi:10.1089/neu.2009.1053

29. Vandevord PJ, Bolander R, SajjaVS, Hay K, Bir CA. Mild neuro-trauma indicates a range-specific

pressure response to low levelshock wave exposure. Ann BiomedEng (2012) 40:227–36. doi:10.1007/s10439-011-0420-4

30. Mac Donald CL, Johnson AM,Cooper D, Nelson EC, Werner NJ,Shimony JS, et al. Detection of blast-related traumatic brain injury inU.S. military personnel. N Engl JMed (2011) 364:2091–100. doi:10.1056/NEJMoa1008069

31. Davenport ND, Lim KO, Arm-strong MT, Sponheim SR. Dif-fuse and spatially variable whitematter disruptions are associ-ated with blast-related mild trau-matic brain injury. Neuroimage(2012) 59:2017–24. doi:10.1016/j.neuroimage.2011.10.050

32. Mac Donald C, Johnson A, CooperD, Malone T, Sorrell J, Shimony J, etal. Cerebellar white matter abnor-malities following primary blastinjury in us military personnel.PLoS One (2013) 8:e55823. doi:10.1371/journal.pone.0055823

33. Mayer AR, Ling J, Mannell MV,Gasparovic C, Phillips JP, DoezemaD, et al. A prospective diffu-sion tensor imaging study in mildtraumatic brain injury. Neurol-ogy (2010) 74:643–50. doi:10.1212/WNL.0b013e3181d0ccdd

34. Ling JM, Pena A, Yeo RA, MeridethFL, Klimaj S, Gasparovic C,et al. Biomarkers of increaseddiffusion anisotropy in semi-acute mild traumatic braininjury: a longitudinal perspec-tive. Brain (2012) 135:1281–92.doi:10.1093/brain/aws073

35. Shah A, Stemper BD, Pintar FA.Development and characterizationof an open-ended shock tube forthe study of blast mTBI. Biomed SciInstrum (2012) 48:393–400.

36. Saatman KE, Contreras PC, SmithDH, Raghupathi R, McDermott KL,Fernandez SC, et al. Insulin-likegrowth factor-1 (IGF-1) improvesboth neurological motor and cogni-tive outcome following experimen-tal brain injury. Exp Neurol (1997)147:418–27. doi:10.1006/exnr.1997.6629

37. Cheney JA, Brown AL, Bareyre FM,Russ AB, Weisser JD, Ensinger HA,et al. The novel compound LOE 908attenuates acute neuromotor dys-function but not cognitive impair-ment or cortical tissue loss follow-ing traumatic brain injury in rats. JNeurotrauma (2000) 17:83–91. doi:10.1089/neu.2000.17.83

38. Zhang H, Avants BB, YushkevichPA, Woo JH, Wang S, McCluskeyLF, et al. High-dimensional spatialnormalization of diffusion tensor

www.frontiersin.org October 2013 | Volume 4 | Article 154 | 11

Budde et al. Blast TBI in the rat

images improves the detection ofwhite matter differences: an exam-ple study using amyotrophic lat-eral sclerosis. IEEE Trans MedImaging (2007) 26:1585–97. doi:10.1109/TMI.2007.906784

39. Newlander SM, Chu A, Sinha US,Lu PH, Bartzokis G. Methodolog-ical improvements in voxel-basedanalysis of diffusion tensor images:applications to study the impactof apolipoprotein E on white mat-ter integrity. J Magn Reson Imaging(2013). doi:10.1002/jmri.24157

40. Wood GW, Panzer MB, Shridha-rani JK, Matthews KA, Capehart BP,Myers BS, et al. Attenuation of blastpressure behind ballistic protectivevests. Inj Prev (2013) 19:19–25. doi:10.1136/injuryprev-2011-040277

41. Nelson TJ, Clark T, Stedje-LarsenET, Lewis CT, Grueskin JM, EcholsEL, et al. Close proximity blastinjury patterns from improvisedexplosive devices in Iraq: a report of18 cases. J Trauma (2008) 65:212–7.

42. Bass CR, Panzer MB, Rafaels KA,Wood G, Shridharani J, Capehart B.Brain injuries from blast. Ann Bio-med Eng (2012) 40:185–202. doi:10.1007/s10439-011-0424-0

43. Fijalkowski RJ, Stemper BD, Pin-tar FA, Yoganandan N, Crowe MJ,Gennarelli TA. New rat model fordiffuse brain injury using coronalplane angular acceleration. J Neuro-trauma (2007) 24:1387–98. doi:10.1089/neu.2007.0268

44. Kwon SK, Kovesdi E, Gyorgy AB,Wingo D, Kamnaksh A, Walker J,et al. Stress and traumatic braininjury: a behavioral, proteomics,and histological study. Front Neurol(2011) 2:12. doi:10.3389/fneur.2011.00012

45. Kamnaksh A, Kovesdi E, Kwon SK,Wingo D,Ahmed F, Grunberg NE, etal. Factors affecting blast traumaticbrain injury. J Neurotrauma (2011)28:2145–53. doi:10.1089/neu.2011.1983

46. Elder GA, Dorr NP, De Gasperi R,Gama Sosa MA, Shaughness MC,Maudlin-Jeronimo E, et al. Blastexposure induces post-traumaticstress disorder-related traits ina rat model of mild traumaticbrain injury. J Neurotrauma (2012)29:2564–75. doi:10.1089/neu.2012.2510

47. Reger ML, Poulos AM,Buen F, Giza CC, Hovda DA,Fanselow MS. Concussive braininjury enhances fear learn-ing and excitatory processesin the amygdala. Biol Psy-chiatry (2012) 71:335–43.doi:10.1016/j.biopsych.2011.11.007

48. Panzer MB, Myers BS, CapehartBP, Bass CR. Development of afinite element model for blastbrain injury and the effects ofCSF cavitation. Ann Biomed Eng(2012) 40:1530–44. doi:10.1007/s10439-012-0519-2

49. Rubovitch V, Ten-Bosch M, ZoharO, Harrison CR, Tempel-BramiC, Stein E, et al. A mousemodel of blast-induced mild trau-matic brain injury. Exp Neurol(2011) 232:280–9. doi:10.1016/j.expneurol.2011.09.018

50. Garman RH, Jenkins LW, SwitzerRC III, Bauman RA, Tong LC,Swauger PV, et al. Blast expo-sure in rats with body shieldingis characterized primarily by dif-fuse axonal injury. J Neurotrauma(2011) 28:947–59. doi:10.1089/neu.2010.1540

51. Mao JC, Pace E, Pierozynski P, KouZ, Shen Y, Vandevord P, et al. Blast-induced tinnitus and hearing loss inrats: behavioral and imaging assays.J Neurotrauma (2012) 29:430–44.doi:10.1089/neu.2011.1934

52. Lipton ML, Kim N, Park YK,Hulkower MB, Gardin TM, ShiftehK, et al. Robust detection oftraumatic axonal injury in indi-vidual mild traumatic brain injurypatients: intersubject variation,change over time and bidirec-tional changes in anisotropy. BrainImaging Behav (2012) 6:329–42.doi:10.1007/s11682-012-9175-2

53. Mac Donald CL, Dikranian K,Bayly P, Holtzman D, Brody D.Diffusion tensor imaging reli-ably detects experimental traumaticaxonal injury and indicates approx-imate time of injury. J Neurosci(2007) 27:11869–76. doi:10.1523/JNEUROSCI.3647-07.2007

54. Chu Z, Wilde EA, Hunter JV,McCauley SR, Bigler ED, Troyan-skaya M, et al. Voxel-based analy-sis of diffusion tensor imaging inmild traumatic brain injury in ado-lescents. AJNR Am J Neuroradiol(2010) 31:340–6. doi:10.3174/ajnr.A1806

55. Budde MD, Janes L, Gold E,Turtzo LC, Frank JA. The contri-bution of gliosis to diffusion ten-sor anisotropy and tractography fol-lowing traumatic brain injury: val-idation in the rat using Fourieranalysis of stained tissue sections.Brain (2011) 134:2248–60. doi:10.1093/brain/awr161

56. Kim JH, Trinkaus K, Ozcan A,Budde MD, Song SK. Postmortemdelay does not change regional dif-fusion anisotropy characteristics in

mouse spinal cord white matter.NMR Biomed (2007) 20:352–9. doi:10.1002/nbm.1138

57. Chavko M, Watanabe T, AdeebS, Lankasky J, Ahlers ST, McCar-ron RM. Relationship between ori-entation to a blast and pressurewave propagation inside the ratbrain. J Neurosci Methods (2011)195:61–6. doi:10.1016/j.jneumeth.2010.11.019

Conflict of Interest Statement: Theauthors declare that the research wasconducted in the absence of any com-mercial or financial relationships thatcould be construed as a potential con-flict of interest.

Received: 31 May 2013; paper pendingpublished: 04 July 2013; accepted: 21 Sep-tember 2013; published online: 14 Octo-ber 2013.Citation: Budde MD, Shah A, McCreaM, Cullinan WE, Pintar FA and StemperBD (2013) Primary blast traumatic braininjury in the rat: relating diffusion ten-sor imaging and behavior. Front. Neurol.4:154. doi: 10.3389/fneur.2013.00154This article was submitted to Neuro-trauma, a section of the journal Frontiersin Neurology.Copyright © 2013 Budde, Shah, McCrea,Cullinan, Pintar and Stemper. This isan open-access article distributed underthe terms of the Creative CommonsAttribution License (CC BY). The use,distribution or reproduction in otherforums is permitted, provided the origi-nal author(s) or licensor are credited andthat the original publication in this jour-nal is cited, in accordance with acceptedacademic practice. No use, distribution orreproduction is permitted which does notcomply with these terms.

Frontiers in Neurology | Neurotrauma October 2013 | Volume 4 | Article 154 | 12